Spatial distribution of early thickness changes in the inner retinal layer in adolescents with Type 1 diabetes

Abstract

Purpose: Diabetic retinopathy (DR) is clinically defined by vascular lesions that develop in the multilayered retina. By the time vascular lesions appear and are clinically detectable, there may already be irreversible damage and permanent vision loss. Inner retinal layer (IRL) thickness changes precede the development of vascular lesions and may serve as an early biomarker of DR. It is unknown whether IRL thickness changes are evenly distributed or localized to discrete regions across the retina.<br /> <br /> We hypothesize that in diabetes, IRL thickness changes are localized to specific regions. The objective of this prospective, cross-sectional observational study was to determine if there are regional changes in IRL thickness in adolescents with Type 1 diabetes compared with controls.

Methods: Participants were adolescents with Type 1 diabetes (T1D) with no or minimal DR (patient group), and adolescents with no diabetes with healthy eyes (control group). At the time of testing, blood glucose levels were maintained within 4-10mmol/L. Standard eye examination and fundus photography ensured adherence to inclusion criteria.<br /> <br /> Using spectral-domain optical coherence tomography (OCT), we obtained high resolution cross-section images of the retina in vivo. A 6x6mm area centered on the fovea was imaged using Cirrus HD-OCT 5000, (Carl Zeiss Meditec). Automated segmentation software was applied to unprocessed OCT images to discriminate individual retina layers. The distance between the upper and lower boundaries of the IRL gives the thickness at that location.

Results: A total of 65536 evenly distributed IRL thicknesses were obtained per participant, achieving up to 11.7 and 2.0μm resolutions in the lateral and axial dimensions, respectively. Currently, single eyes from 5 patients (mean age 23.0±1.7y) and 8 controls (mean age 23.4±4.7y) have been analyzed. Natural variation was observed in each participant. Analysis revealed no significant difference in total IRL thickness between patient and control groups although the patient group showed greater variance. Regionally, patient foveas were thicker, superior perifoveas thinner, and inferior perifoveas thicker, than controls.

Conclusions: This is the first study to maintain high resolution segmentation of retinal layer thickness throughout the analyses. These data show certain regions that are prone to changes in diabetes.